Quantification of Variability and Uncertainty for Selected Nonroad Mobile Source Emission Factors

Emission factors are subject both to variability and uncertainty. Variability refers to real differences in emissions among multiple emission sources at any given time or over time for any individual emission source. Uncertainty refers to lack of knowledge regarding the true value of emissions. In this paper, we demonstrate quantitative methods for characterizing both variability and uncertainty and apply the methods to case studies of emission factors for Lawn and Garden (L&G) engines and for construction, farm, and industrial (CFI) equipment. Data were obtained from emissions testing conducted by others. Databases were created and statistically analyzed to determine the minimum number of emission source categories with statistically significantly different average emissions. Inter-engine variability in emissions was quantified using parametric distributions. Bootstrap simulation was used to characterize confidence intervals for the fitted distributions. For 2-stroke L&G engines, the 95 percent confidence intervals for the mean emission factors for total hydrocarbon (THC) and Nitrogen Oxides (NOx) emissions in g/hp-hr units were -32% to +38% and -46% to +65%, respectively. For 4-stroke L&G engines, the confidence intervals for mean emissions in g/hp-hr units were -38% to +45% for THC and -25% to +38% for NOx. For CFI engines, which are primarily diesel, the 95 % confidence intervals for the mean emission factors were as small as –10% to +11% and as large as –48% to +49%. These quantitative measures of uncertainty convey information regarding the quality of the emission factors and serve as a basis for calculation of uncertainty in emission inventories. The method, example case studies, and benefits of the approach are discussed.